Article
Food Tracking Perspective: DNA Metabarcoding to
Identify Plant Composition in Complex and
Processed Food Products
Antonia Bruno1 , Anna Sandionigi1,* , Giulia Agostinetto1 , Lorenzo Bernabovi1,†, Jessica Frigerio2 , Maurizio Casiraghi1 and Massimo Labra1
1 Zooplantlab, Department of Biotechnology and Biosciences, University of Milano-Bicocca, Piazza della
Scienza 2, I-20126 Milano, Italy; antonia.bruno@unimib.it (A.B.); g.agostinetto@campus.unimib.it (G.A.); andrea.galimberti@unimib.it (L.B.); maurizio.casiraghi@unimib.it (M.C.); massimo.labra@unimib.it (M.L.)
2 FEM2-Ambiente, Piazza della Scienza 2, I-20126 Milano, Italy; jessica.frigerio@fem2ambiente.com
* Correspondence: anna.sandionigi@unimib.it; Tel.: +39-02-64483413
† This paper is dedicated to the memory of our wonderful colleague, Dr. Lorenzo Bernabovi, who recently
passed away.
Received: 9 February 2019; Accepted: 19 March 2019; Published: 25 March 2019
Abstract: One of the main goals of the quality control evaluation is to identify contaminants in raw material, or contamination after a food is processed and before it is placed on the market. During the treatment processes, contamination, both accidental and economically motivated, can generate incongruence between declared and real composition. In our study, we evaluated if DNA metabarcoding is a suitable tool for unveiling the composition of processed food, when it contains small trace amounts. We tested this method on different types of commercial plant products by using tnrL marker and we applied amplicon-based high-throughput sequencing techniques to identify plant components in different food products. Our results showed that DNA metabarcoding can be an effective approach for food traceability in different type of processed food. Indeed, the vast majority of our samples, we identified the species composition as the labels reported. Although some critical issues still exist, mostly deriving from the starting composition (i.e., variable complexity in taxa composition) of the sample itself and the different processing level (i.e., high or low DNA degradation), our data confirmed the potential of the DNA metabarcoding approach also in quantitative analyses for food composition quality control.
Keywords: DNA metabarcoding; trnL; molecular markers, High Throughput Sequencing; food tracking; food contaminants; processed food; food matrices; herbal products
1. Introduction
One of the main goals of food quality control is to identify contaminants or counterfeits in raw material, during food processing and before market placement [1]. During treatment processes, contamination, both accidental or economically motivated, can lead to incongruences between declared and real composition of the final food products. These incongruences could be threats to consumers health and may reduce potential health benefits of the products, even causing allergic or toxic reactions [2]. A faster, more efficient and cost-effective quality check can solve these issues and promotes the import and export of verified food products to different countries [3,4]. Together with the production of novel foods, methods for the development of new products and new production strategies (i.e., the choice of ingredients or components, the design of food processing, shelf-life studies and sensory evaluation of products) are steadily changing equilibria between the two forces. Consequently, the analysis and control of food content are constantly being developed. After recent
technological advanced, molecular techniques, based on the amplification and/or sequencing of marker DNA regions, can be a useful diagnostic method to identify food species composition and represents one of the the most promising tools [5].
The use of DNA-based markers in molecular authentication allows accurate and sensitive results, even in cases of phylogenetically related species [6]. This is proved by the collaboration among FDA’s Center for Food Safety and Applied Nutrition (CFSAN) and Center for Veterinary Medicine (CVM), the University of Guelph’s Biodiversity Institute of Ontario, Canadian Center for DNA Barcoding, and the Laboratories of Analytical Biology at the Smithsonian National Museum of Natural History, Suitland, MD, USA. A coordinated network of these institutes resulted in a universal and standardized approach for metazoan identification based on DNA barcoding: they first provided details on protocols, reagents, and equipment required to carry out a validation study for barcode generation for fish identification in the form of FDA Laboratory Information Bulletin No. 4420. Although the molecular approach is well supported and validated in the case of the identification of a single biological raw material [7–10], other issues arise when dealing with treated or adulterated foods. In fact, highly processed products are one of the challenges of food traceability because DNA could undergo degradation processes due to the intense physico-chemical conditions of industrial treatments. As a consequence, in several processed foods only very short fragments of DNA are available for the analysis and typically the common DNA barcoding analysis fails. In this context, High-Throughput DNA Sequencing (HTS) techniques (also called Next Generation DNA Sequencing, NGS), combined with powerful tools for data analysis, are revolutionizing food quality control. HTS can be applied using an amplicons-based strategy (known as DNA metabarcoding) or a genomes-based strategy (known as metagenomics). In particular, DNA metabarcoding is presently a new standard in the quality control of new food products [11–13]. Nevertheless, DNA metabarcoding still has some limitations: (i) DNA fragmentation [14,15]; (ii) DNA amplification biases [16,17]; (iii) the real ‘primer universality’ [18]; (iv) the presence of DNA amplification inhibitors [19]; (v) the occurrence of (food) materials in trace and low DNA yield [17]; (vi) the accidental laboratory contamination when the target DNA is fragmented and of low concentration. A sequence variation due to these restrains hampers species detection in food matrices.In particular when a false mutation derived to degenerative process occurs, the assignment of the incorrect query sequence returns false positive due a misidentified assignment [20].
The intron fragment P6-loop (10–143 bp) of the chloroplast tRNALeu (UAA) intron sequences [tnrL (UAA), 254–767 bp] is a short, but informative sequence proved by Taberlet et al. [21] to be effective in identifying plant species in processed food and ancient permafrost samples. Despite closely related species are not always resolved [21], food industry [22], forensic sciences [23,24] and diet studies based on faeces [25] successfully used tnrL marker for plant identification and this contributed to increment the number of tnrL sequences present in databases. In our study, we evaluated if DNA metabarcoding is a suitable tool for unveiling the composition of processed food, even when it contains trace amounts. We tested this method on different types of commercial plant products by using the tnrL marker.
2. Materials and Methods
2.1. Commercial Processed Foods and Mock Herbal Mixture for Qualitative Analysis
To test the efficacy of DNA metabarcoding, we chose six different commercial products (described in Figure1) with the following characteristics: (i) ingredients not easily distinguishable by morphology; (ii) specified ingredients in the label; (iii) different processes involved in the production (i.e., deep-freezing, lyophilization). As representative of these features, we selected from the market processed food such as vegetable stock cube, curry, pureed soup of deep-frozen vegetables, or food supplements, that are characterized by high levels of food processing and, at the same time, high level of complexity considering taxa composition. In particular, considering processing, vegetable stock cube undergoes to high temperature processing and dehydration. Curry is a complex combination of spices or herbs, usually including ground turmeric, cumin, coriander, ginger, processed in the form of powder. Pureed soup of deep-frozen vegetables undergoes both to high temperature processing and deep-freezing for storage. Food supplements can be manufactured using intact sources, phytoextracts or other not specified fractions from plants, and can be in the form of liquid, pills, capsules or tablets, as in our case study. We also selected products that are not characterized by a high level of complexity considering taxa composition, but can be subjected to adulteration, such as flavoured tea and saffron. In particular, saffron adulteration often occurs due to the high cost of the raw material. Commonly, adulteration consist in adding or mixing of similar materials such as beet, pomegranate, red-dyed silk fibers, safflower, and marigold or honey and vegetable oils [26]. Moreover, a mixture of plants, mimicking commercial herbal products, known in composition and quantity (100 mg in dry weight for each specimen), was created by a herbalist’s shop.
2.2. Fruit Mixtures Known in Composition and Quantity for Quantitative Analysis
To test the efficacy of DNA metabarcoding analysis in identifying the composition of commercial products at the qualitative and quantitative level, we prepared five mixtures composed by five plants commonly used to prepare fruit juice and with consistent phylogenetic distances: Ananas (Ananas comosus Bromeliaceae), Avocado (Persea americana Lauraceae), Dragon fruit (Hylocereus undatus Cactaceae), Mango (Mangifera indica Anacardiaceae), Papaya (Carica papaya Caricaceae). To avoid possible laboratory contaminants, fruit not present in the ingredients declared in the selected commercial food and never processed in our laboratory were chosen. To avoid possible intra-species sequence variability, we used only one fruit of each type to create the artificial mixtures. The five mixtures were prepared by mixing different amounts of DNA of each species (Table S1), in different combinations: in detail, DNA was individually extracted from each fruit and individually quantified using qPCR, as described in Section2.3and in [27]. Then, we prepared different dilutions of each DNA extract. Finally, we composed the artificial mixtures starting from the DNA extracts and combining equal volumes but different concentrations of DNA, as reported in Table S2. As a consequence, we obtained five mixtures all composed by the same five fruits, but the combination of DNA fruit concentrations is peculiar of each mixture (Table S1). The combination of different DNA quantities instead of fruit sample quantities expressed as weight was preferred in order to evaluate whether the number of features obtained is proportional to the DNA copy numbers of amplified trnL fragments. 2.3. DNA Extraction and qPCR
qPCR assays were performed with AB 7500 (Applied Biosystems, Carlsbad, CA, USA) on DNA extracts deriving from commercial processed foods, mock herbal mixture, each single fruit used for the artificial fruit mixtures and the artificial fruit mixtures themselves. qPCR conditions included an initial denaturation at 95◦C for 100, followed by 40 cycles of denaturation at 95◦C for 1500and annealing-elongation for 10at 55◦C. Real-time PCR was set up with 2X SsoFast EvaGreen Supermix with Low ROX (Bio-Rad S.r.l., Segrate (MI), Italy) in which EvaGreen was used as a detecting dye; a 10 µL reaction consisted of 5.0 µL SsoFast EvaGreen Supermix with Low ROX, 0.1 µL each 10 µmol/L primer solution, 2 µL DNA sample, and 2.8 µL of Milli-Q water. Primer sequences are given in Table1. All samples and negative controls were run in triplicate. Amplification data were collected and analyzed with the SDS 7500 Real-Time PCR System Software (Applied Biosystems). All the procedures were conducted in the laminar flow cabinet, to avoid contamination with exogenous DNA and inter-samples contamination and in separate rooms for the pre and post amplification steps, with dedicated personal protective equipment (PPE).
Table 1.Primer sequences used for Illumina sequencing (trnL g-h) and for Sanger sequencing (trnL c-h).
trnL g GGGCAATCCTGAGCCAA
trnL h CCATTGAGTCTCTGCACCTATC
trnL c CGAAATCGGTAGACGCTACG
2.4. Sanger Sequencing
The trnL region amplified by the primer pairs trnL g-h (the forward primer position is upstream from trnL c) [21] was sequenced in the samples belonging to exotic fruit used for artificial fruit mixtures, in order to obtain the reference sequences implementing our local reference database. PCR amplification was performed using KAPA HiFi HotStart Ready Mix (Kapa Biosystems, Woburn, MA, USA) in a 25 µL reaction volume, according to the manufacturer’s instructions. PCR cycles consisted of an initial denaturation for 10 min at 94◦C, 35 cycles of denaturation (3000at 94◦C), annealing (3000 at 55◦C), elongation (6000at 72◦C), and final elongation at 72◦C for 60. Primer sequences are given in Table1. PCR products were purified from agarose gel with QIAquick Gel Extraction Kit (Qiagen, Hilden, Germany), following manufacturer’s protocol. Products were submitted for sequencing to Eurofins Genomics (https://www.eurofinsgenomics.eu/). The DNA strands were bidirectionally sequenced. All obtained sequences were manually edited, aligned, and used as query in an individual Basic Local Alignment Search Tool (BLAST) in GenBank to verify samples identity.
2.5. Libraries Preparation and Sequencing
2.6. Bioinformatics Analysis
DNA metabarcoding analysis of food composition was performed using the plugins of the QIIME2 suite [28]). Raw Illumina reads were paired and pre-processed using VSEARCH v2.5.0—merge pairs algorithm [29]. Reads were filtered out if ambiguous bases were detected and were shorter than 52 bp. Moreover, an expected error = 1 was used as an indicator of read accuracy [30]. Plant features were obtained using—cluster_ f ast algorithm with a 100% sequence identity with at least a depth of 500×
for each feature. A random sequence was chosen as the representative sequence of the cluster. To decrease the false positive rate in the sequence population, a chimera detection analysis was performed on the obtained reference sequences. Since there is no reference database for trnL gene for chimera detection, we used uchime_denovo algorithm that carries out a de novo analysis without a reference. The taxonomic assignment of the representative sequences, to obtain the features, was carried out using the classi f y-consensus-vsearch plugin implemented in QIIME2 [31] against the local database adopting a consensus confidence threshold of 0.9. Reference feature sequences were aligned against our local reference database, built with downloaded trnL sequence available in NCBI at November 2018 and increased with Sanger sequences obtained in our laboratory for fruit mixture samples as a control. Where consensus confidence during classification was equal to 1 the query sequence was assigned to the corresponding species. An additional comparison with the non-redundant nucleotide NCBI database was performed in order to check the features that were not assigned by our reference database. To estimate differences among the replicates of each fruit mix, we performed beta diversity analysis. Bray-curtis distance was calculated and one-way PERMANOVA pairwise test was performed using beta-group-signi f icance plugin implemented in QIIME2 platform [28]. To assess the accuracy with which the expected taxonomic composition is reconstructed after experimental analysis, we used the q2-quality-control plugin implemented in QIIME2. Linear regression scores between expected and observed feature abundances (qPCR copies vs reads amount) are calculated at species rank, and plots of accuracy and observation correlations are plotted.
3. Results and Discussion 3.1. Sequence Analysis
To asses the food composition in the six selected processed foods (n = 6) and in forty-one mock admixtures (n = 41), 6,308,509 quality-filtered trnL sequences were obtained. After filtering and primer removal, the remaining sequences were of high quality and had an average length of 75 bp (ranging from 32 to 84 bp) and clustered into 671 features (100% identity).
3.2. DNA Metabarcoding for Food Traceability
We focused on the DNA metabarcoding analysis of the six commercial products to estimate their composition and to compare it with the list of ingredients accompanying the selected products. The biological diversity observed in the commercial food is characterized by 22 taxa assigned, 16 of which at species level. If we assume the declared composition on the label as the true composition, the rate of false positive (misclassified features or real contaminants) occurred (n = 9) is lower than the false negative (not sequenced features) (n = 18). Given the different type of food processing and consequent different level of DNA degradation occurred for each product, we now discuss case by case the results obtained, resumed in Table1and illustrated in Figure1.
Saffron
Crocus sativus*.
Flavoured tea
Camellia sinensis*. Natural flavours: Citrus sinensis, Syzygium aromaticum, Cinnamomum verum.
Echinacea purpurea*, Betula pendula*, Centella asiatica*, Trigonella foenum-graecum*, Chamerion angustifolium*, Foeniculum vulgare*, Taraxacum officinale*, Viola tricolor*.
Mock herbal mixture
25% 100%
Foeniculum vulgare, Crocus sativus, Streptophyta unassigned, Laurus sp., Medicago sativa. Vegetal stock cube
Allium cepa*, Daucus carota*, Petroselinum crispum*, Solanum tuberosum*, Allium ampeloprasum*, Solanum lycopersicum*, Apium graveolens, Spinacia oleracea*, Allium sativum*, Spices.
90%
Pisum sativum, Foeniculum vulgare, Camellia sinensis. Pureed soup of deep-frozen vegetables
Daucus carota*, Cucurbita pepo, Solanum lycopersicum*, Brassica oleracea*, Cucurbita sp., Solanum tuberosum*, Petroselinum crispum, Spinacia oleracea*, Allium ampeloprasum, Allium cepa, Ocimum basilicum.
50%
Medicago sativa. Food supplement
Matricaria chamomilla, Gentiana lutea, Achillea millefolium, Curcuma longa, Taraxacum officinale*, Lamium album, Peumus boldus, Foeniculum vulgare, Mentha piperita, Origanum majorana, Cnicus benedictus.
9% 100%
Daucus carota. Curry
Piper nigrum, Cuminum cyminum, Coriandrum sativum*,
Cinnamomum verum, Curcuma longa, Syzygium aromaticum, Myristica fragrans, Trigonella foenum-graecum*, Capsicum sp., Brassica*, Allium sp.*.
36%
Figure 1.Schematic overview of commercial processed foods and mock herbal mixture tested with DNA metabarcoding approach. The percentage bars report the “observed/expected” percentage, that is the match in what was declared in the ingredients label and what we detected using DNA metabarcoding; not declared specimen/contaminants are excluded from the percentage calculation
and listed in the boxes on the right. See also Table1.
Rosa caninaCentella asia
ca Glycine max Magnoliophyta sp. Petroselinum sp. Streptophyta unassigned Medicago sa va Laurus sp.
Chamerion angusfolium
Echinacea sp. Coriandrum savum Betula pendula Crocus savus Camellia sinensis Foeniculum vulgare Allium sp. Taraxacum ocinale Spinacia oleracea Daucus carota Trigonella foenum-graecum Viola tricolor Brassica oleracea Solanum sp. Pisum savum Deep-froz enveg etable Mock herbal mixtur e Veg etable stock cube Food supplemen t
Figure 2.Taxa barplot reporting relative abundances of plants detected in processed food.
and to trace food contaminants. Nevertheless, DNA metabarcoding in plants does have limitations, mainly arising from the combination of difficulty in DNA amplification due to degraded DNA in the processed samples and the primer bias in the annealing step.
Table 2.List of processed food and artificial lab mix tested. When percentages were displayed in the ingredients label, the information was reported in the “declared composition” column.
DNA Metabarcoding Not Declared Specimen Sample Declared Species Composition Composition (Vsearch) Contaminants
(>0.3%) (False Positive)
saffron Crocus sativus Crocus sativus (99.8%)
flavoured tea
Camellia sinensis Camellia sinensis (99.7%) natural flavours:
Citrus sinensis Syzygium aromaticum
Cinnamomum verum
vegetable stock cube
Allium cepa (1.9%) Allium sp. (39.3%) Foeniculum vulgare (43%) Daucus carota (1%) Spinacia oleracea (9.3%) Crocus sativus (2.1%) Petroselinum crispum (0.5%) Daucus carota (3.9%) Streptophyta unassigned (0.5%)
Solanum tuberosum (0.3%) Petroselinum sp. (0.3%) Laurus sp. (0.5%) Allium ampeloprasum (0.2%) Solanum sp. (0.3%) Medicago sativa (0.3%) Solanum lycopersicum (0.1%) Apium graveolens (0.1%) Spinacia oleracea (0.1%) Allium sativum (0.1%) spices curry
Piper nigrum Trigonella foenum-graecum (36%) Daucus carota (13.4%) Cuminum cyminum Brassica sp. (33.3%)
Coriandrum sativum Coriandrum sativum (13.2%) Cinnamomum verum Allium sp. (4%)
Curcuma longa Syzygium aromaticum Myristica fragrans Trigonella foenum-graecum Capsicum sp. Brassica sp. Allium sp. deep-frozen vegetables
Daucus carota Spinacia oleracea (37.2%) Pisum sativum (24.8%) Cucurbita pepo Solanum sp. (13.3%) Foeniculum vulgare (19.8%) Solanum lycopersicum Daucus carota (2.6%) Camellia sinensis (1.1%)
Brassica oleracea Brassica oleracea (0.9%) Cucurbita sp. Solanum tuberosum Petroselinum crispum Spinacia oleracea Allium ampeloprasum Allium cepa Ocimum basilicum food supplement
Matricaria chamomilla Taraxacum officinale (99.2%) Medicago sativa (0.4%) Gentiana lutea Achillea millefolium Curcuma longa Taraxacum officinale Lamium album Peumus boldus Foeniculum vulgare Mentha piperita Origanum majorana Cnicus benedictus
mock herbal mixture
Echinacea purpurea Taraxacum officinale (29.3%) Betula pendula Viola tricolor (26.3%) Centella asiatica Betula pendula (12.8%) Trigonella foenum-graecum Echinacea sp. (9.9%)
Chamerion angustifolium Foeniculum vulgare (8.6%) Foeniculum vulgare Chamerion angustifolium (6%) Taraxacum officinale Trigonella foenum-graecum (5.8%)
3.3. DNA Metabarcoding to Quantitatively Evaluate Food Composition
As already noted in our results on processed food (see in the previous paragraph the case of Taraxacum officinale), amplicon metabarcoding can be affected by the PCR amplification step using “universal” markers. The occurrence of a bias during PCR amplification may cause the inaccurate estimation of quantities and this was at least partially demonstrated for metazoan and plants [17,35,36]. This bias generates a variable number of template–primer mismatches across species, resulting in a final amplified DNA mixture that does not always reflect the original proportion of each species, limiting the quantitative potential of DNA metabarcoding [37–40]. Assuming probability of primer bias, we have been concerned about whether there is a correlation between DNA copies obtained through qPCR amplification and reads obtained by sequencing, for each tested species. This could shed a light on the evaluation of food composition also from a quantitative point of view, especially in the case of samples heterogeneous in composition. For this reason, we set up an assay where five different fruit mixtures were composed by five different exotic fruits, in different known proportions. We opted for fruit mixtures mimicking exotic fruit juices because fruit is relatively easy to authenticate by their morphological characteristics when intact and fresh, but give rise to the possibility of adulteration when processed into juice. In particular, adulteration may regard both composition and proportions of the fruit, preferring preferring cheaper and more available fruit. We decided to create our own ad hoc mixture without using any commercial fruit mixture, since in this way we could be sure that all the DNA sequences belonged to same fruit (for each fruit mixture), and there was no intra-species sequence variability. Fruit mixtures were analysed both with qPCR and Illumina HiSeq sequencing, using the same primer pairs. As shown in Figure3, we were able to detect and correctly assign each exotic fruit used for mixture preparation, also thanks to our local reference database implemented with Sanger sequences generated for this experiment.
r1 r2 r3 r4 r5 r6 r7 r8 Exp r1 r2 r3 r4 r5 r6 r7 r8 Exp r1 r2 r3 r4 r5 r6 r7 r8 Exp r1 r2 r3 r4 r5 r6 r7 r8 Exp r1 r2 r3 r4 r5 r6 r7 r8 Exp
Ananas sp.
Persea americana
Hylocereus undatus
Mangifera indica
Carica papaya
Unassigned
100%
75%
50%
25%
0%
Contaminants
Streptophyta
Mix 1
Mix 2
Mix 3
Mix 4
Mix 5
The replicates of each fruit mix showed a high repeatability of the obtained results, confirming the accuracy of metabarcoding method (The Bray-Curtis distance among each samples are reported in Table S5). Analyzing the relationship between expected and observed quantities of fruit (Figure4), we verified that a strong correlation existed in the case of mix 2 (r = 0.9978, p-value < 0.0001) , mix 4 (r = 0.9586, p-value = 0.0006) and mix 5 (r = 0.9945, p-value < 0.0001). On the contrary, the relationship was not significative for mix 1 (r = 0.6041, p-value > 0.05) and mix 3 (r = 0.3631, p value > 0.05). This bias could be due to primer competition in a heterogeneous solution, where the final DNA concentration will reflect the initial one when there is no bias in primer affinity for all species; otherwise, the initial and final DNA concentrations would be poorly related. Considering the results reported here, when the analysis rely only on dry weight, the quantitative potential of the technique is partial, but qualitative results (the species list) are reliable. This depends on the fact that same weights do not necessarily carry with them (i) the same number of cells and (ii) the same number of trnL copies. On the other hand, when DNA metabarcoding analysis is put in relation to DNA copies, a quantitative correlation can be estimated if primer competition is low or can be measured. In general, a well standardized protocol, from experimental design to bioinformatic analysis, including internal controls and DNA quantification, will maximize the power of DNA metabarcoding.
0.25 0.00 0.25 0.50 0.75 1.00 1.25 Expected abundance 0.25 0.00 0.25 0.50 0.75 1.00 1.25 Observed abundance
Figure 4.Plots of accuracy and observation correlations for each samples in the five fruit mixtures.
4. Conclusions
In recent years, we face the boosting of new food processes, developed using various tools of science, engineering, and biotechnology. This leads to the placement on the market of food products that are modified, improved, and made to look better, taste different, and be commercially attractive, generating an increase in global trading of such food products and food supplements. Together with the advances in food processing, we are facing also the challenge of new food adulteration and counterfeits strategies and, as a consequence, an accurate authentication is essential (i) for correct labeling, (ii) to ensure food safety, (iii) to protect the consumer, (iv) to sustain the economic value of high quality, controlled, food. At the same time, there is a boosting in molecular techniques development and application, and an exponential increase of molecular data available in GenBank database, paving the way for the overcoming of the caveat and pitfalls demonstrated and discussed in this research.
application will require a shared and standardized protocol, which is still lacking despite the fast progresses obtained till now. Today, technical aspects limitations are challenging our effort to obtain an optimal tool for both qualitative and quantitative analyses. Nevertheless, considering the exponential advancement in scientific research and molecular approaches, we are confident that in the next few years we will be able to apply in an extensive, rapid and cost-effective way these molecular tools starting from complex and degraded matrices.
Supplementary Materials:The following are available online athttp://www.mdpi.com/2073-4425/10/3/248/s1, Table S1: Amounts of DNA of each species in the five different mix of mock herbal mixture; Table S2: different concentrations of DNA, based on qPCR quantifications of each mix in mock herbal mixture; Table S3: List of barcodes used for Illumina sequencing; Table S4: Bray-curtis distance matrix of fruit mixture samples; Table S5: OTU table of mixed fruit with taxonomic annotation; Table S6: OTU table of preocessed food with taxonomic annotation; Table S7: Regression results of mixed fruit comparison.
Author Contributions:Conceptualization, A.B., A.S., M.C. and M.L.; Data curation, A.S.; Formal analysis, A.S.; Funding acquisition, M.C. and M.L.; Methodology, A.B., A.S., G.A., L.B. and J.F.; Project administration, M.C. and M.L.; Resources, M.C. and M.L.; Software, A.S., G.A. and L.B.; Supervision, M.C. and M.L.; Validation, A.B., M.C. and M.L.; Visualization, A.S. and G.A.; Writing—original draft, A.B., A.S., M.C. and M.L.; Writing—review & editing, A.B., A.S., M.C. and M.L.
Funding:This work was partially supported by Regione Lombardia in the framework of the Project “Food Social Sensor Network—Call Accordi per la Ricerca e l’Innovazione” co-funded by the Italian Government action “POR FESR 2014–2020”. Barcode sequences were kindly provided by Francesco G. Ficetola.
Acknowledgments: This manuscript is dedicated to Lorenzo Bernabovi, for his kindness and his enthusiasm. All the ZooPlantLab is grateful to have had him as a colleague and friend, his joy will always be remembered.
Conflicts of Interest:The authors declare no conflict of interest.
References
1. Valentini, P.; Galimberti, A.; Mezzasalma, V.; De Mattia, F.; Casiraghi, M.; Labra, M.; Pompa, P.P. DNA
Barcoding Meets Nanotechnology: Development of a Universal Colorimetric Test for Food Authentication.
Angew. Chem. Int. Ed. 2017, 56, 8094–8098. [CrossRef]
2. Spencer, P.; Berman, F. Plant toxins and human health. In Food Safety: Contaminants and Toxins;
Oxford University Press: Oxford, UK; CABI Publishing: Wallingford, Oxfordshire, UK, 2003; pp. 1–23.
3. Barling, D.; Sharpe, R.; Lang, T. Traceability and ethical concerns in the UK wheat—Bread chain: From food
safety to provenance to transparency. Int. J. Agric. Sustain. 2009, 7, 261–278. [CrossRef]
4. Charlebois, S.; Sterling, B.; Haratifar, S.; Naing, S.K. Comparison of global food traceability regulations and
requirements. Compr. Rev. Food Sci. Food Saf. 2014, 13, 1104–1123. [CrossRef]
5. Lo, Y.T.; Shaw, P.C. DNA-based techniques for authentication of processed food and food supplements.
Food Chem. 2018, 240, 767–774. [CrossRef]
6. De Mattia, F.; Bruni, I.; Galimberti, A.; Cattaneo, F.; Casiraghi, M.; Labra, M. A comparative study of
different DNA barcoding markers for the identification of some members of Lamiacaea. Food Res. Int. 2011,
44, 693–702. [CrossRef]
7. Galimberti, A.; Sandionigi, A.; Bruno, A.; Bruni, I.; Barbuto, M.; Casiraghi, M.; Labra, M. Towards a
universal molecular approach for the quality control of new foodstuffs. In Advances in Food Biotechnology; Wiley Blackwell: New York, NY, USA; 2015; p. 37.
8. Bruni, I.; Galimberti, A.; Caridi, L.; Scaccabarozzi, D.; De Mattia, F.; Casiraghi, M.; Labra, M. A DNA
barcoding approach to identify plant species in multiflower honey. Food Chem. 2015, 170, 308–315. [CrossRef]
9. Barbuto, M.; Galimberti, A.; Ferri, E.; Labra, M.; Malandra, R.; Galli, P.; Casiraghi, M. DNA barcoding
reveals fraudulent substitutions in shark seafood products: The Italian case of “palombo” (Mustelus spp.).
Food Res. Int. 2010, 43, 376–381. [CrossRef]
10. de Boer, H.J.; Ichim, M.C.; Newmaster, S.G. DNA barcoding and pharmacovigilance of herbal medicines.
Drug Saf. 2015, 38, 611–620. [CrossRef]
11. Galimberti, A.; Bruno, A.; Mezzasalma, V.; De Mattia, F.; Bruni, I.; Labra, M. Emerging DNA-based
12. Staats, M.; Arulandhu, A.J.; Gravendeel, B.; Holst-Jensen, A.; Scholtens, I.; Peelen, T.; Prins, T.W.; Kok, E. Advances in DNA metabarcoding for food and wildlife forensic species identification. Anal. Bioanal. Chem.
2016, 408, 4615–4630. [CrossRef]
13. Raclariu, A.C.; Heinrich, M.; Ichim, M.C.; de Boer, H. Benefits and limitations of DNA barcoding and
metabarcoding in herbal product authentication. Phytochem. Anal. 2018, 29, 123–128. [CrossRef]
14. Hrnˇcírová, Z.; Bergerová, E.; Siekel, P. Effects of technological treatment on DNA degradation in selected
food matrices of plant origin. J. Food Nutr. Res. 2008, 47, 23–28.
15. Gryson, N. Effect of food processing on plant DNA degradation and PCR-based GMO analysis: A review.
Anal. Bioanal. Chem. 2010, 396, 2003–2022. [CrossRef]
16. Shokralla, S.; Spall, J.L.; Gibson, J.F.; Hajibabaei, M. Next-generation sequencing technologies for environmental
DNA research. Mol. Ecol. 2012, 21, 1794–1805. [CrossRef]
17. Elbrecht, V.; Leese, F. Can DNA-based ecosystem assessments quantify species abundance? Testing primer
bias and biomass—Sequence relationships with an innovative metabarcoding protocol. PLoS ONE 2015,
10, e0130324. [CrossRef]
18. Deagle, B.E.; Jarman, S.N.; Coissac, E.; Pompanon, F.; Taberlet, P. DNA metabarcoding and the cytochrome c
oxidase subunit I marker: Not a perfect match. Biol. Lett. 2014, 10, 20140562. [CrossRef]
19. Schrader, C.; Schielke, A.; Ellerbroek, L.; Johne, R. PCR inhibitors–occurrence, properties and removal.
J. Appl. Microbiol. 2012, 113, 1014–1026. [CrossRef]
20. Ficetola, G.F.; Taberlet, P.; Coissac, E. How to limit false positives in environmental DNA and metabarcoding?
Mol. Ecol. Resour. 2016, 16, 604–607. [CrossRef]
21. Taberlet, P.; Coissac, E.; Pompanon, F.; Gielly, L.; Miquel, C.; Valentini, A.; Vermat, T.; Corthier, G.;
Brochmann, C.; Willerslev, E. Power and limitations of the chloroplast trn L (UAA) intron for plant
DNA barcoding. Nucleic Acids Res. 2006, 35, e14. [CrossRef]
22. Faria, M.; Magalhães, A.; Nunes, M.; Oliveira, M. High resolution melting of trnL amplicons in fruit juices
authentication. Food Control 2013, 33, 136–141. [CrossRef]
23. Yang, Y.; Xie, B.; Yan, J. Application of next-generation sequencing technology in forensic science.
Genom. Proteom. Bioinform. 2014, 12, 190–197. [CrossRef]
24. Young, J.M.; Weyrich, L.S.; Cooper, A. Forensic soil DNA analysis using high-throughput sequencing:
A comparison of four molecular markers. Forensic Sci. Int. Genet. 2014, 13, 176–184. [CrossRef]
25. Kartzinel, T.R.; Chen, P.A.; Coverdale, T.C.; Erickson, D.L.; Kress, W.J.; Kuzmina, M.L.; Rubenstein, D.I.;
Wang, W.; Pringle, R.M. DNA metabarcoding illuminates dietary niche partitioning by African large
herbivores. Proc. Natl. Acad. Sci. USA 2015, 112, 8019–8024. [CrossRef]
26. Heidarbeigi, K.; Mohtasebi, S.S.; Foroughirad, A.; Ghasemi-Varnamkhasti, M.; Rafiee, S.; Rezaei, K. Detection
of adulteration in saffron samples using electronic nose. Int. J. Food Prop. 2015, 18, 1391–1401. [CrossRef]
27. Bruno, A.; Sandionigi, A.; Galimberti, A.; Siani, E.; Labra, M.; Cocuzza, C.; Ferri, E.; Casiraghi, M. One step
forwards for the routine use of high-throughput DNA sequencing in environmental monitoring: An efficient and standardizable method to maximize the detection of environmental bacteria. MicrobiologyOpen 2017,
6, e00421. [CrossRef]
28. Bolyen, E.; Rideout, J.R.; Dillon, M.R.; Bokulich, N.A.; Abnet, C.; Al-Ghalith, G.A.; Alexander, H.; Alm, E.J.;
Arumugam, M.; Asnicar, F.; et al. QIIME 2: Reproducible, Interactive, Scalable, and Extensible Microbiome
Data Science. PeerJ Preprints 2018. [CrossRef]
29. Rognes, T.; Flouri, T.; Nichols, B.; Quince, C.; Mahé, F. VSEARCH: A versatile open source tool for metagenomics.
PeerJ 2016, 4, e2584. [CrossRef]
30. Edgar, R.C.; Flyvbjerg, H. Error filtering, pair assembly and error correction for next-generation sequencing
reads. Bioinformatics 2015, 31, 3476–3482. [CrossRef]
31. Bokulich, N.A.; Kaehler, B.D.; Rideout, J.R.; Dillon, M.; Bolyen, E.; Knight, R.; Huttley, G.A.; Caporaso, J.G.
Optimizing taxonomic classification of marker-gene amplicon sequences with QIIME 2’s q2-feature-classifier
plugin. Microbiome 2018, 6, 90. [CrossRef]
32. Särkinen, T.; Staats, M.; Richardson, J.E.; Cowan, R.S.; Bakker, F.T. How to open the treasure chest?
Optimising DNA extraction from herbarium specimens. PLoS ONE 2012, 7, e43808. [CrossRef]
33. Sahu, S.K.; Thangaraj, M.; Kathiresan, K. DNA extraction protocol for plants with high levels of secondary
34. Burns, M.; Wiseman, G.; Knight, A.; Bramley, P.; Foster, L.; Rollinson, S.; Damant, A.; Primrose, S. Measurement issues associated with quantitative molecular biology analysis of complex food matrices
for the detection of food fraud. Analyst 2016, 141, 45–61. [CrossRef]
35. Balech, B.; Sandionigi, A.; Manzari, C.; Trucchi, E.; Tullo, A.; Licciulli, F.; Grillo, G.; Sbisà, E.; De Felici, S.;
Saccone, C.; et al. Tackling critical parameters in metazoan meta-barcoding experiments: A preliminary
study based on coxI DNA barcode. PeerJ 2018, 6, e4845. [CrossRef]
36. Thudi, M.; Li, Y.; Jackson, S.A.; May, G.D.; Varshney, R.K. Current state-of-art of sequencing technologies for
plant genomics research. Brief. Funct. Genom. 2012, 11, 3–11. [CrossRef]
37. Piñol, J.; Senar, M.A.; Symondson, W.O. The choice of universal primers and the characteristics of the species
mixture determine when DNA metabarcoding can be quantitative. Mol. Ecol. 2019, 28, 407–419. [CrossRef]
38. Bista, I.; Carvalho, G.R.; Tang, M.; Walsh, K.; Zhou, X.; Hajibabaei, M.; Shokralla, S.; Seymour, M.; Bradley, D.;
Liu, S.; et al. Performance of amplicon and shotgun sequencing for accurate biomass estimation in
invertebrate community samples. Mol. Ecol. Resour. 2018, 18, 1020–1034. [CrossRef]
39. Elbrecht, V.; Vamos, E.E.; Steinke, D.; Leese, F. Estimating intraspecific genetic diversity from community
DNA metabarcoding data. PeerJ 2018, 6, e4644. [CrossRef]
40. Leray, M.; Yang, J.Y.; Meyer, C.P.; Mills, S.C.; Agudelo, N.; Ranwez, V.; Boehm, J.T.; Machida, R.J. A new
versatile primer set targeting a short fragment of the mitochondrial COI region for metabarcoding metazoan
diversity: Application for characterizing coral reef fish gut contents. Front. Zool. 2013, 10, 34. [CrossRef]
c
